The Icelandic Financial Supervisory Authority would do well to tread cautiously when blaming the Iceland Stock Exchange (part of NASDAQ-OMX) for deficient systems. In 2006, I was part of the first big data enterprise in Iceland which was based on a common metrics database. The system consisted of a common metrics database and XML relays, statistical functions and charting tools and looked like this (click on images to enlarge):
The Monitor charting component with capability to run any time series against any other. Common metrics settings automatically adjusted frequency, scale and currency and offered users the ability to aggregate series of higher frequency to match those with lower frequency (e.g. monthly to yearly or daily to monthly).
The charting panel could be expanded to display 1,000 charts distributed across multiple monitors; we ran this on 8 monitors at a bank preceding a sale to that same bank. A click on the chart could save it as image, extract the underlying data, send as email or apply global functions such as inflation or growth rates at a single click. If the FSA had adopted it, it would have been able to track the entire economy far more efficiently that was done at the time and even today.
The charts could be saved as reports to be autoloaded whenever desired. Watching numerous charts autoload with information from dozens of sources with functions and settings was like magic and far beyond anything the FSA had at the time or has today. This is what the institution was offered; this is what it rejected and now attempts to throw dust in the air in order to free itself from taking responsbility for allowing the country to crash when it was preventable.
Flipping the panels into tabular view yielded underlying series including derivative data and functions. The first panel of this screenshot shows raw data as appearing on the chart, then an organized table (ideal to monitor financial reports). The matrix autogenerated various standard indicators such as percentage change from last period, accumulated growth rate and so on.
The correlation matrix is the key component. With it, it was possible to measure the relationship between stock prices, volume, inflation, population, GDP, forex and anything else contained within the database. That component alone showed beyond any doubt March 2006 that the Icelandic banks were lending so excessively that a crash was nearly inevitable. That was why the FSA was approached and the reception was bizarre to say the least. The outcome: Total collapse of the Icelandic financial system and the FSA refuses to take responsibility for it.
No other system at the time was capable to deliver metadata within the charting framework. This was yet another leading edge that could have been used to train personnel and deliver information to the market and government to improve risk management and the national level.
A multilingual dictionary made it possible to toggle between several languages which is something we’re still not seeing on the market today. Metadata tables were easy to update and could be saved, printed, stored or otherwise shared. Series without metadata are risky as it is necessary to know exactly how the data is derived and what it means.
Again, this component was part of the complete system the FSA imagined it could build internally using taxpayers money. That decision led to a shockwave of bankruptcies at corporate and household levels and cost the country billions. If Parliament would launch an inquiry into this, it would find that Iceland did not have to collapse at all. At the time we did not have a name for this kind of system, now, 6 years later, they are known as ‘Big Data’.
We had built a database that contained time series covering the entire Icelandic economy and was used to project the eventual crash of the banks as early as March 2006. This was around the same time we were in talks with the FSA regarding the importance of having such a system in-house to monitor financial activity and detect emerging threats.
To outline the inner mechanics of the system:
- A uniform data structure made it easy to run any time series data against any other. Frequency, scale, currency and even language were made uniform and usable with any other application developed by third parties. This means that comparing financial activity of all financial corporations was not only possible but nearly automated. Monitoring stock activity against financial performance; piece of cake.
- Built-in functions made it possible to run any kind of statistical analysis on the series, but the heavyweight was the massive correlation engine that showed the relationships between all time series in the database. That in itself packed more power back in 2006 than the FSA has today or will have in the near future. FSA is still NOT handling the data management issue and is as backwards now as it was in 2006. Therefore, I have no trust in that organization whatsoever.
- Data organization and interaction was structured to clean up the update process and conform the haphazard tables used by the FSA, where pension fund, insurance company and banking financials are manually updated using Excel tables. That data had all been pooled into a uniform database and could be viewed, plotted, extracted or relayed in any way desired. Uniformity also meant that sophisticated model building was greatly simplified as builders did not have to concern themselves with preparing the data – it was all ready to go.
- The architecture allowed extensions to other institutions in the form of data relays, where updated take place. These relays could be manually fed or hardwired using XML interfaces. We successfully applied it to the Murex trading platform where it performed error free and required no maintenance. The same would have been the case for the entire Icelandic government and would have prevented (or at least minimized the effect of) the collapse. Information is power and if the information spells crash, both regulator and government are forced to react.
The system was introduced at the FSA early 2006 and well received, until suddenly all doors shut. There was no explanation other than that the FSA were going to develop its own solution. And now that same organization tried to blame the stock exchange for own negligence. The blame for the collapse of the Icelandic credit system lies with FSA and that can be proved. One of us went before a Parliament commission with the case of public information management competing with private sector development since it was apparent that the FSA and all other institutions were completely ignorant of the dangerous financial position the country was in and adamant to remain so. The lack of transparency and resistance from leading agencies (Central Bank, Ministry of Finance and Statistics Iceland included) blindfolded regulators and made the government unaware that Iceland was on the brink of financial disaster. As already stated, this was known in March 2006. Watching Iceland move closer and closer to the brink of destruction and knowing that public institutions didn’t care was torture. Seeing those same individuals still in their positions and paid by taxpayers is the equivalent of rubbing salt in the wound.
When the Icelandic economy collapsed, I assisted the Special Investigative Commission with the bank reports since that data was nowhere to be found. Not even the FSA had this information which says a lot about how that operation stood in terms of information management. Today, the FSA is trying to whitewash itself and attempting to cast blame on the stock exchange. That simply will not be tolerated and I’m prepared to go before any Parliament commission and drive the nails home that will pin the FSA firmly to the wall. The FSA was offered the system they needed at the time but decided to develop it in-house. Therefore, it should be held liable for the damage caused by that decision and the entire management.
ADDITION DECEMBER 28
A report like that could be saved for instant retrieval, including functions such as those that appear in chart 6 (% change). This was introduced to key government institutions, which could have seamlessly connected the entire economy from root level (e.g. banks) to aggregates, something these same institutions are incapable of doing today, 5 years later.